The long range goal of this project is to identify common genetic variants that predispose patients to alkylator-associated AML. Therapy-related acute myeloid leukemia (t-AML) is a lethal, increasingly prevalent complication of alkylator chemotherapy. Several lines of evidence suggest that there is a genetic component to t-AML susceptibility. However, common predisposing genetic factors for t-AML have not yet been identified. We hypothesize that germline variants of genes somatically mutated in de novo AML ("leukemia pathway" genes) may be susceptibility alleles for t-AML. The relevant genetic variants may include small insertions, deletions, single nucleotide polymorphisms (SNPs), or larger segmental DMA copy number variants (CNVs). We will utilize complementary approaches in mice and humans to identify candidate susceptibility alleles and validate their importance for human t-AML. This Project capitalizes on several strengths of GAML members. In collaboration with Project 2, we have developed expertise in the identification of CNVs using data from high-density oligonucleotide array-based comparative genomic hybridization experiments. In collaboration with Projects 2, 4 and Core D, we are identifying novel variants in "leukemia pathway" genes that we are incorporating into gene association studies. During the first funding period, we identified several mouse strains susceptible or resistant to alkylator-associated t-AML. Haplotype association mapping in 20 parental strains implicated Mlf1 as a candidate t-AML susceptibility factor. Congenic Mlfl null strains are being generated to validate this candidate susceptibility gene. We have accrued subjects and acquired specimens for genotyping from a large number of patients with alkylatorassociated t-AML and matched controls from our institution and outside collaborators. We propose the following two aims that build on these initial findings to identify the genetic basis of alkylator-associated t- AML: Specific Aim 1. We will identify genetic factors influencing t-AML susceptibility in mice. We will perform a whole genome scan to screen comprehensively for t-AML susceptibility loci in mice. We will map CNVs in t-AML susceptible and resistant strains and correlate CNV profiles with t-AML susceptibility. High priority candidate susceptibility factors discovered in Aims 1 and 2 will be validated using mouse models. Specific Aim 2. We will define the importance of germline genetic variants for human t-AML susceptibility. Using a case-control design, we will use both candidate gene and whole genome approaches to determine the impact of SNPs and CNVs on risk of alkylator-associated t-AML in humans. Identification of predisposing genetic factors should lead to personalized therapies with reduced risk of t- AML.